Introduction

The antibiotic pressure in some hospital departments, especially Intensive Care Units (ICUs), the inappropriate use of antibiotics, and the difficulty in detecting certain resistance phenotypes have led to a major worldwide increase in bacterial resistance rates [1] and favored the development of multiple resistances in microorganisms such as Staphylococcus spp. [2]. In this genus, the association of methicillin-resistance with resistance to aminoglycosides, fluoroquinolones, tetracyclines, and macrolides, among other antibiotics, is well documented [3].

There have been increasing reports of methicillin-resistant staphylococci, mainly S. aureus (MRSA), with reduced susceptibility to glycopeptides (glycopeptide-intermediate-resistant S. aureus, GISA) [4] and even of isolates resistant to vancomycin and teicoplanin (glycopeptide-resistant S. aureus, GRSA) [5]. This emerging phenomenon has also been described in other species of coagulase-negative staphylococci [6]. Hence, it is especially important to correctly detect isolates with heterogeneous resistance to glycopeptides (heteroresistant glycopeptide intermediate S. aureus, hGISA) [7]. Critically, hGISA isolates might be precursors of GISA strains [8], associated with therapeutic failure [9]. All of these developments have limited the success of hospital therapies, increasing the morbidity and mortality of infected patients [10].

This situation has prompted the introduction of novel antibiotics against multiresistant Gram-positive microorganisms, e.g., quinupristin-dalfopristin, linezolid, tigecycline, and daptomycin, which have proven effective against GISA, hGISA, and GRSA isolates [11, 12].

Ever since linezolid was approved for the treatment of Gram-positive bacteria in our country in 2001, its pharmacokinetic and pharmacodynamic properties have led to its increasing use for the main indications, i.e., nosocomial pneumonia, community-acquired pneumonia, and complicated skin and soft tissue infections [13]. It inhibits the synthesis of proteins by binding to the domain V region of the 23S rRNA gene [14]. Mutations in this domain have been associated with resistance to linezolid in such species as S. aureus, coagulase-negative staphylococci, Enterococcus faecalis, and E. faecium [15–17]. G2576T mutation is the most frequent cause of resistance to linezolid found in clinical isolates of Enterococcus spp. and Staphylococcus spp. [18]. Other reported mutations include G2447T and T2500A in S. aureus, C2534T in S. epidermidis, G2505A in E. faecium, and C2512T, G2513T, and C2610G in E. faecalis [19, 20]. Resistance to linezolid is also produced by the action of Cfr-rRNA methyltransferase, which is known to confer resistance to phenicols, lincosamides, pleuromutilins, and streptogramin A [21].

We report the presence of various methicillin-resistant isolates of S. hominis subsp. hominis with resistance to linezolid and a reduced susceptibility to glycopeptides in blood cultures from ICU or Emergency Area patients at our third-level university hospital in Southern Spain.

Materials and methods

Bacterial isolates and patients

We studied ten clinical isolates of S. hominis subsp. hominis from nine patients (eight in the ICU and one in the Emergency Area) admitted to our hospital between November 2007 and March 2008. In all cases, S. hominis subsp. hominis was isolated in blood cultures (BacT/ALERT 3D; bioMérieux, Marcy-l’Etoile, France). Isolates were identified by using the WIDER system (Francisco Soria Melguizo, Madrid, Spain), confirming the result by 16S rRNA gene sequencing.

Data were gathered (see Table 1) on the date of ICU (or Emergency Area) admission, date of positive blood cultures, number of blood cultures in which S. hominis was identified with respect to the total number of blood cultures, diagnosis causing ICU/Emergency Area admission, antibiotic treatments before and since admission, and the presence of other microorganisms obtained from different samples, when applicable.

Table 1 Profile of the nine patients with the presence of Staphylococcus hominis in the blood culture
Full size table

In six patients (In Table 1, P4 to P9), S. hominis was isolated in only one blood culture out of multiple serial extractions during their hospital stay. It was isolated from a blood culture taken from one of these patients (patient P9) in the Emergency Area but not in subsequent blood cultures taken in the ICU. S. hominis was isolated in two blood cultures taken from patient P3 on different days (isolates S-6911 and S-7063). S. hominis was isolated in three blood cultures taken on the same day from patient P1 and in four blood cultures taken on the same day from patient P2. However, only one isolate from each of these two patients could be phenotypically and genotypically analyzed. The others were not preserved because they were the first two patients and the clinical significance of S. hominis had not yet been recognized.

Antimicrobial susceptibility testing

Susceptibility of the isolates to antibiotics commonly used against staphylococci (Table 2) was measured by Etest (AB Biodisk, Solna, Sweden) with 0.5 McFarland inoculum in Mueller-Hinton agar (bioMérieux), following the manufacturer’s recommendations. Clinical and Laboratory Standards Institute (CLSI) [22] criteria were applied to define the cutoff points for all antibiotics except for tigecycline. This was assigned a minimum inhibitory concentration (MIC) cutoff value of ≤0.5 µg/ml, in accordance with recommendations of the European Committee on Antimicrobial Susceptibility Testing (EUCAST) [23].

Table 2 Minimum inhibitory concentration (MIC) values (in µg/ml) obtained by Etest in Mueller-Hinton agar at 24 h of incubation at 37°C for the antibiotics assayed in the ten clinical isolates of Staphylococcus hominis
Full size table

In isolates with MIC values ≥4 µg/ml for vancomycin or ≥8 µg/ml for teicoplanin, the Etest macromethod was used [24] to detect any heterogeneous resistance to glycopeptides [7]. Briefly, 100 µl of an inoculum suspension to 2.0 McFarland was pipetted onto a standard BHI agar plate (bioMérieux). Etest vancomycin and teicoplanin MIC gradient strips (AB Biodisk) were then applied, incubating the plates at 35°C, and reading them at 24 and 48 h, taking the result at 48 h as the endpoint. An isolate was considered to show heterogeneous resistance to glycopeptides when its MIC value was ≥8 µg/ml for vancomycin and ≥8 µg/ml for teicoplanin or ≥12 µg/ml for teicoplanin [7].

Following CLSI recommendations [22], S. aureus ATCC 29213 was used as the control in all susceptibility studies.

PCR amplification of the domain V region and DNA sequencing

Domain V region of the 23S rRNA gene spanning 2011 to 2699 bp (E. coli numbering) was amplified. Oligonucleotide primers 5′-TGGGCACTGTCTCAACGA and 5′-ATCCCGGTCCTCTCGTACTA were used to amplify a 690-bp fragment. Polymerase chain reaction (PCR) conditions were 5 min at 95°C; 35 cycles at 95°C (1 min), 55°C (1 min), and 72°C (2 min); and a final 5 min at 72°C. Amplified DNA products were resolved by electrophoresis in agarose (1% wt/vol) gels containing ethidium bromide. Finally, the PCR product was purified (Wizard PCR Preps DNA Purification System; Promega, Madison, WI, USA), processed with a DNA sequencing kit, and analyzed in an automatic DNA sequencer (Applied Biosystems, Foster City, CA, USA). DNA sequences obtained were aligned with the corresponding nucleotide sequences from a linezolid-susceptible S. aureus reference strain (GenBank accession number X68425).

PFGE

We used a modification of the protocol of Murchan et al. [25], carrying out two procedures, digesting agarose blocks with SmaI and then with ApaI, and loading them into two agarose gels at 1% wt/vol. Electrophoresis was done in 0.5X Tris-Borate-EDTA buffer in a CHEF-DRII apparatus (Bio-Rad Laboratories, Hemel Hempstead, U.K.) under the following conditions: pulses of 0.1 to 30 s, 6 V/cm for 24 h at a temperature of 12 to 14°C. The gels were stained with ethidium bromide, visualized with ultraviolet light, and photographed. Fragment patterns were interpreted as described by Tenover et al. [26].

Results

Table 2 lists the MIC values obtained for the ten isolates in the antibiotic susceptibility study using Etest in Mueller-Hinton agar. All were resistant to oxacillin (MIC > 256 µg/ml), gentamicin (MIC ≥ 16 µg/ml), levofloxacin (MIC ≥ 4 µg/ml), cotrimoxazole (MIC > 32/608 µg/ml), and linezolid (MIC ≥ 24 µg/ml) (Fig. 1a), and susceptible to daptomycin (MIC ≤ 0.25 µg/ml) and tigecycline (MIC ≤ 0.38 µg/ml). Nine of the isolates were resistant to erythromycin (MIC > 256 µg/ml) and with intermediate resistance to clindamycin (MIC = 2 µg/ml), whereas the remaining isolate (S-7063) was susceptible to both antibiotics.

Fig. 1
figure 1

Etest results for linezolid (a: suspension to 0.5 McFarland in Mueller-Hinton agar at 24 h of incubation) and for vancomycin and teicoplanin (b and c: suspension to 0.5 McFarland in Mueller-Hinton agar at 24 h of incubation; d and e: suspension to 2.0 McFarland in BHI agar after 48 h of incubation) in one of the studied isolates of Staphylococcus hominis (S-855)

Full size image

All isolates except for S-7063 showed an MIC value of 2 µg/ml for vancomycin and of 12 µg/ml for teicoplanin in the Etest in agar Mueller-Hinton (Fig. 1b, c). Subsequent application of the Etest macromethod in BHI yielded MIC values of 8 µg/ml for vancomycin and 12 µg/ml for teicoplanin (Fig. 1d, e), confirming the presence of strains with heterogeneous resistance to vancomycin in these nine isolates.

Figure 2 depicts amplification of the 690-bp fragment of the domain V region of the 23S rRNA gene, performed in all isolates. Subsequent sequencing of this fragment revealed the presence of C2190T and G2603T mutations in nine of the isolates (S-6567, S-6639, S-6911, S-117, S-127, S-532, S-650, S-855, and S-1018) and of C2190T and G2474T mutations in isolate S-7063 (Fig. 3).

Fig. 2
figure 2

Detection of the 690-bp fragment of domain V of the 23S rRNA gene by agarose gel electrophoresis (lane 1, molecular weight marker VIII, Roche Diagnostics, Spain)

Full size image
Fig. 3
figure 3

Mutations detected in the studied isolates of Staphylococcus hominis

Full size image

PFGE with SmaI digestion (Fig. 4) showed that all isolates except for S-7063 had the same electrophoretic profile, showing only three bands. A subsequent PFGE with ApaI digestion (Fig. 5) showed a pattern for these nine isolates that confirmed their epidemiologic relationship. Isolates in lines 1 to 6 and line 8 showed the same profile (profile A) and isolates in lines 7 and 9 were considered as two subtypes: profile A1 and A2, respectively. Isolate S-7063 in line 10 showed a completely different profile (profile B).

Fig. 4
figure 4

Pulsed-field gel electrophoresis (PFGE) of ten Staphylococcus hominis digesting with SmaI

Full size image
Fig. 5
figure 5

PFGE of ten Staphylococcus hominis digesting with ApaI

Full size image

Discussion

Coagulase-negative staphylococci, which can form part of normal skin flora, are also a frequent cause of infections in hospitalized patients. They are commonly implicated in catheter infections, infectious endocarditis, prosthesis infections, and osteoarticular infections, among others. Their treatment with glycopeptides or linezolid, especially in ICUs, has considerably increased the consumption of these antibiotics in recent years. As with other families of antibiotics, it has been reported that the increase in their use is related to an increase in the detection of resistant isolates [19, 27].

This is the first case since the introduction of glycopeptides and linezolid in our hospital of a circulating strain of methicillin-resistant Staphylococcus spp. with resistance to linezolid and a reduced susceptibility to glycopeptides.

The previous administration of linezolid [28] or glycopeptides [24] is known to be associated with the development of resistance in coagulase-negative staphylococci. There have also been reports of infection or colonization by linezolid-resistant coagulase-negative staphylococci in patients with no previous exposure to the antibiotic [28] and of environmental contamination by methicillin-resistant staphylococci with reduced susceptibility to glycopeptides [29].

In the majority of our patients, the bacterium was identified in only one of several serial blood cultures; therefore, this microorganism can be considered as a contaminant of these blood cultures rather than being responsible for a true bacteremia. Only one patient (P2) showed the microorganism in four blood cultures after presenting with nosocomial pneumonia at several days after ICU admission for medication overdose, but it was not detected in the sputum cultures from this patient.

One possible explanation is that these bacteria were transmitted via the hands of the healthcare staff or were present in the ICU environment [30]. However, prevention measures recommended for the management of MRSA-colonized patients were taken [31], and none of the ICU staff or patients screened for nasal carriage were colonized with the resistant strain and no environmental contamination was detected. The origin of the outbreak remains unknown.

Various mutations in the domain V region of the 23S rRNA gene have been associated with resistance to linezolid in staphylococci, but only G2576T and G2603T mutations has been described in clinical isolates [18, 32]. G2576T mutation was not detected in our isolates. However, we report, for the first time, the presence of two mutations simultaneously in linezolid-resistant Staphylococcus spp.

The clinical significance of this circulating strain was not initially understood, since S. hominis was isolated with other microorganisms frequently implicated in nosocomial infections and, therefore, is considered a contaminant, and because S. hominis was isolated in only one culture from some of these patients.

Hospitalized patients, especially ICU patients, can be infected by colonized patients or healthcare personnel. The possible selection over time of isolates of the more virulent S. aureus that are MRSA, GISA, and/or resistant to linezolid poses a potentially severe challenge to current anti-Gram-positive therapies. Although the prevention of dissemination is important, we believe that the rational use of antibiotics in preventing the selection of drug-resistant pathogens is as important.